2
$\begingroup$

I'm looking for Jupyter extension to plot interactive graphs. For example I need to plot twenty time series lines with order to examine data. Only plugin which I've found is bqplot https://github.com/bloomberg/bqplot but its documentation is too complicated.

Could you advice some other free tool or help with bqplot?

I can't plot lines with different colors. 

unitIds = commonSeries['UnitId'].unique()
commonSeries = commonSeries.sort_values(by=['Date','UnitId'], ascending=True)
commonSeries = commonSeries.set_index(pd.DatetimeIndex(commonSeries['Date']))

plt.clear()
x_ord = OrdinalScale()
y_sc = LinearScale()
lines = []

for unitId in unitIds:    
    timeseries = commonSeries.loc[commonSeries['UnitId'] == unitId]
    line = plt.Lines(x=timeseries.index, y=timeseries['Value'], scales={'x': x_ord, 'y': y_sc}, colors=['red','green','blue'])
    lines.append(line)

plt.Figure(marks=lines)

enter image description here

Improve this question
$\endgroup$

4 Answers 4

Reset to default
3
$\begingroup$

Plotly is by far the best interactive visualization library/platform I have used, and it works very well with IPython/Jupyter too.

There are tutorials on Plotly's docs which help you integrate it with Jupyter. Another tutorial.

Improve this answer
$\endgroup$
1
$\begingroup$

I've found simple, useful and opensource plugin: http://holoviews.org

Improve this answer
$\endgroup$
1
$\begingroup$

Bokeh and Plotly both feature interactive visualizations and can be used in a Jupyter notebook.

Improve this answer
$\endgroup$
1
$\begingroup$

Simply use %matplotlib notebook the same way you would use %matplotlib inline.

https://www.reddit.com/r/IPython/comments/36p360/try_matplotlib_notebook_for_interactive_plots/

Improve this answer
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.